Biophysics at the dawn of exascale computers
Biophysics at the dawn of exascale computers
复制标题
百亿亿次计算机初期的生物物理学
DOI：
10.1016/j.bpj.2023.06.017
复制
发表时间：
2023
影响因子：
3.4
通讯作者：
Chipot, Chris
中科院分区：
文献类型：
--
作者：
Singharoy, Abhishek;Pérez, Alberto;Chipot, Chris
我们很高兴为纪念2022年5月在德国汉堡举行的BPS专题会议，推出题为“百亿亿次计算机黎明的生物物理学”的特刊。通过会谈和广泛的讨论，分子和细胞生物学家、化学家、物理学家、数学家和计算机科学家努力寻找共享创新和讨论未来需求的共同点，并作为一个社区向前发展，利用前沿资源。这种跨学科的讨论通过选择地点得到加强，即欧洲x射线自由电子激光设施的地点，在那里光物质相互作用的物理学被利用来学习从分子到细胞和组织的生物学。这期特刊集中体现了会议的多学科精神，以及16篇论文中对生物物理计算的整体多尺度观点，将实验家和理论家聚集在一起，思考百亿亿次计算将在生物物理领域实现的令人兴奋的可能性。该问题涵盖了广泛的生物系统，从简单的蛋白质和核酸到用于分子识别和运动的寡聚复合物。与此同时，现实溶剂化和膜模型的方法论发展已经被报道，探测了大量的时空尺度（介观到原子），同时融合了生物物理、计算和定量实验的见解。我们汇集了在蛋白质折叠和组装、变构途径解剖、大分子相互作用和细胞自下而上结构等广泛领域工作的实验学家和理论家，在这些领域，大规模计算有望带来重大发现。现在着重介绍一些典型的调查结果。计算生物物理学有助于确定可能的作用机制，否则仅靠实验很难确定。这种协同作用可以在Khandelia及其同事(1)的工作中看到，他们将蛋白质结合对膜曲率的影响合理化，并利用这一知识来设计增加结合的新序列。访问百亿亿次计算将进一步加强实验和计算社区之间的串扰(2)。Gumbart和他的同事提出了一项新的、值得注意的计算工作，使用增强的采样工具来调和与已知实验数据兼容的perpern折叠机制(3)。蛋白质展开研究使用加速分子动力学模拟与自适应方法(4)。通过将分子线索与表型结果连接起来，分子动力学模拟的参数用于建立抗生素耐药性起源研究的布朗动力学和群体动力学(5)和错义多态性的作用(6)。溶剂化(7)和核酸在这次会议上也有代表，研究范围从原子力场的重要性，以表示它们的结构和动力学(8)到更粗粒度的水平，以了解分子凝聚物的形成(9)或染色质纤维的行为（10）。计算还有助于解开复杂分子过程背后的机制，从小底物穿过膜的渗透（11）到FOF1-ATP合成酶中复杂的催化旋转步骤（12）。尽管百亿亿次计算有望突破目前分子动力学模拟的限制，最终在现实的时间尺度上解决细胞器到细胞尺度的系统，但对非常大的…
We are happy to present the special issue titled Biophysics at the Dawn of Exascale Computers to commemorate the BPS Thematic Meeting held in Hamburg, Germany, in May 2022. Through the talks and extensive discussions, molecular and cellular biologists, chemists, physicists, mathematicians, and computer scientists worked to find common ground about sharing innovations and debate on future needs—and to move forward as a community to take advantage of leading-edge resources. Such interdisciplinarity in the discourse was reinforced by the choice of the venue, namely the site of the European X-Ray Free-Electron Laser Facility, where the physics of light-matter interactions is harnessed to learn biology from molecules up to cells and tissues. The special issue epitomizes this multidisciplinary spirit of the conference and a holistic multiscale view on biophysical computations arranged in 16 contributions, bringing together experimentalists and theoreticians to think about the exciting possibilities that exascale computing will enable in the field of biophysics. The issue covers a breadth of biological systems ranging from simple proteins and nucleic acids to their oligomeric complexes used for molecular recognition and locomotion. Concomitantly, methodological developments of realistic solvation and membrane models have been reported that probe a multitude of spatiotemporal scales (mesoscopic to atomistic), while concomitantly melding biophysical, computational, and quantitative experimental insights. We brought together experimentalists and theoreticians working in the broad areas of protein folding and assembly, dissection of allosteric pathways, macromolecular interactions, and bottom-up structure of cells, wherein large-scale computing is expected to bring forth major discoveries. Some exemplary findings are now highlighted. Computational biophysics helps to identify possible mechanisms of action that would otherwise be difficult to identify by experiments alone. Such synergies can be seen in the work of Khandelia and co-workers (1) in which they rationalize the effect of protein association on membrane curvature and use this knowledge to engineer novel sequences that increase association. Access to exascale computing will further enhance cross talk between experiments and computational communities (2). Gumbart and co-workers present a new and notable computational effort using enhancing sampling tools to reconcile a folding mechanism for pertactin that is compatible with known experimental data (3). Protein unfolding is investigated by using accelerated molecular dynamics simulations with adaptive approaches (4). By bridging molecular cues with phenotypic outcomes, parameters from molecular dynamics simulations are used to set up Brownian and population dynamics of studies on the origins of antibiotic resistance (5) and the role of missense polymorphisms (6). Solvation (7) and nucleic acids were also represented in this conference, with studies ranging from the importance of atomistic force fields to represent their structure and dynamics (8) to the more coarse-grained level to understand how molecular condensates form (9) or how chromatin fibers behave (10). Computations also help untangle the mechanisms that underlie complex molecular processes ranging from the permeation of small substrates across membranes (11) to the intricate catalytic rotary step in FOF1-ATP synthase (12). Although exascale computing offers the promise of pushing back the current limits of molecular dynamics simulations to ultimately tackle organelle-to-cell-scale systems over realistic timescales, investigation of very large …